© Global Warming Images / WWF

Fast-forward to the past

13 June 2019

This article originally appeared in The Circle: Arctic Check-up. The Circle shares perspectives from across the Arctic, and the views expressed here are not necessarily those of WWF. See all Circle issues here.

We often hear about how our rapidly changing climate is already accelerating the melting of glaciers and ice sheets across the globe. But what about the reverse: could the melting polar ice sheets themselves disrupt climate, creating a feedback loop that leads to runaway environmental change? To explore that idea, NICK GOLLEDGE brought together an international team of climate scientists—and the results were alarming.

INSIDE A FROZEN TENT at the end of July 2010, Dorthe Dahl-Jensen, a Danish paleoclimatologist, announced the completion of the North Greenland Eemian Ice Drilling (NEEM) project. The team had finally hit bedrock, 2,537 metres below the surface of the Greenland Ice Sheet. Dahl-Jensen and her team had spent three summers camped high on the summit plateau, painstakingly drilling to recover cores that would tell them how the ice sheet had changed during the Eemian period 125,000 years ago.

This interval—the last time Earth was warmer than it is today—is important because it poses a worrying question. Average air temperatures then were only a degree or so above current temperatures, yet the global sea level was more than six metres higher than it is now. How could this be possible? As we head toward an artificially warmer world, can the NEEM ice core records give us a window into our future?

One way to address this question is to use computer models to simulate these periods of the past, and then use the same models to project into the future. For this we need three things: evidence of how the world (particularly the polar ice sheets) looked during warmer periods; predictions of how the climate may change in the future; and a numerical ice-sheet model that incorporates the necessary physical equations to accurately simulate the flow of an entire ice sheet.

In our recent work, we used a model developed by a team at the University of Alaska, Fairbanks. Over the last 10 years, we have used this model to predict how both the Greenland and Antarctic ice sheets have evolved in the past, as well as how they might change in the future. By checking our past simulations against geological records as well as ice core records from sites such as NEEM, we have developed confidence that the model can be used reliably.


Climate models suggest that our current governmental commitments to mitigating greenhouse gas emissions will lead to average global warming of 3°C to
4°C by the end of the century. When we used our model to simulate how the ice sheets would respond to this warming, we found they lost ice, not just through melting at their surface, but also through dynamic thinning, a process through which ice-sheet outlet glaciers are melted by heat from seawater, thinning their lowest regions. Because this thinning happens most swiftly at the terminus, or end of the glacier, the overall gradient of the ice surface gets steeper, which in turn encourages the ice to flow more quickly and discharge even more of its mass into the ocean.

In Greenland, our model suggested that this would happen to the greatest extent in the northwest sector of the ice sheet. Using a climate model, we then calculated that this meltwater would ultimately find its way into the North Atlantic. And this is where the problems begin: even though the amount of meltwater released is tiny compared with the volume of the ocean, it is buoyant because it contains no salt, so it floats on the sea surface. Normal ocean circulation relies on a process known as convective overturning, in which relatively warm water rises from depth at high latitudes, releases heat to the atmosphere and, once cooler, begins to sink.

But this sinking relies on the water being salty, and therefore dense. Adding freshwater, even if it is cold, prevents this sinking process. As the meltwater from Greenland flows south, it meets warm water from the Gulf of Mexico being carried north. When the Gulf Stream moves east, it releases so much heat into the atmosphere that northwest Europe is kept far warmer than areas at comparable latitudes in eastern Canada. As this current slows down, however, this heat transport reduces, and northwest Europe cools. But the heat in the ocean remains—and it has to go somewhere.


With the rate of sinking suppressed, there is less vertical mixing in the water column, so the warmer, deeper layers retain their heat. This heat is especially dangerous for ice sheets, because many outlet glaciers flow through deep troughs that can be hundreds of metres below sea level and beneath the cold sea surface. Instead, these glaciers are flowing into the deeper, warmer parts of the water column, where the extra heat—trapped by the reduced overturning—accelerates the melting of the submarine ice fronts. The glaciers once again get steeper, flow more quickly and discharge even more ice.

We found that meltwater from the Greenland Ice Sheet can accelerate its own retreat by up to 40 per cent.

These processes form a self-reinforcing loop, also known as positive feedback. Yet the complexity of combining ice sheet models with global climate models has meant that this feedback was ignored in previous predictions. By including this loop, we found that meltwater from the Greenland Ice Sheet can accelerate its own retreat by up to 40 per cent.

Furthermore, the disruption to ocean circulation caused by ice sheet meltwater extends across the Arctic, reducing air temperatures in parts of Siberia and the Aleutian Islands, but producing con- siderably warmer temperatures across Svalbard and as far north as the North Pole. Worse still, the discharge of ice sheet meltwater into the oceans seems to upset circulation patterns in a way that can amplify year-to-year climate variations, producing unreliable weather that can be far warmer or colder than average from one year to the next.

Recent research has shown that the levels of warming expected over the coming centuries have no analogue in the recent past. In fact, conditions by the end of this century will be similar to those last seen three million years ago, when the sea level was around 20 metres higher than it is now. More
alarmingly, some studies have suggested that positive feedbacks could trigger an unstoppable chain of events that would lead to a “hothouse” world entirely different from the one we inhabit today.

Whether we choose to fast-forward into that kind of future is down to us. But evidence from the past suggests that unless we rapidly reduce our emissions, the changes we are setting in motion now may play out for hundreds—or more likely, thousands—of years.

NICK GOLLEDGE is an associate professor in the Antarctic Research Centre at Victoria University of Wellington in New Zealand. He uses computer simulations of the Greenland and Antarctic ice sheets to understand how they behave under different climates and what this might mean for global sea levels.